Authors: Sunil Kumar, Damodar Prasad Dewanda, Dr. Dharm Pal, Bharti Soni
Abstract: Accurate and high-resolution air pollution exposure data are critical for understanding urban health risks, yet traditional monitoring networks often lack the spatial density needed to capture intra-urban variability. This study presents a fine-scale air pollution exposure mapping framework using low-cost sensor networks deployed across a major metropolitan area. By integrating real-time data from particulate matter (PM₂.₅ and PM₁₀), nitrogen dioxide (NO₂), and ozone (O₃) sensors with land-use and meteorological data, we generate detailed spatial-temporal pollution maps at the neighborhood scale. Machine learning models are used to calibrate and validate sensor outputs against regulatory-grade monitors, ensuring reliability. The results reveal significant spatial heterogeneity in pollutant concentrations, strongly correlated with traffic density, industrial zones, and microclimatic factors. Hotspot analysis identifies vulnerable populations disproportionately exposed to harmful air pollutants. This approach demonstrates the potential of low-cost sensor networks for democratizing air quality data, informing urban planning, and advancing environmental justice in rapidly growing cities.
DOI: http://doi.org/10.5281/zenodo.15806733
